The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by design. It is instructed to read and prioritize rules from project-level files such as .ai-factory/DESCRIPTION.md, .ai-factory/ARCHITECTURE.md, and specifically .ai-factory/skill-context/aif-plan/SKILL.md.
Ingestion points: The skill pulls implementation requirements and behavioral overrides from several project files, including config.yaml and roadmap artifacts.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat content from these external files as untrusted or to ignore embedded natural language instructions.
Capability inventory: The skill possesses significant capabilities that could be misused if the input files are compromised, including shell command execution for git and filesystem operations (Bash), and the ability to define tasks for subsequent execution via TaskCreate and subagents.
Sanitization: The instructions do not mandate validation or sanitization of the content extracted from the project artifacts before it is used to influence the agent's logic.

aif-plan